Silas Alberti (@silasalberti) / X

Silas Alberti

753 posts

Silas Alberti

@silasalberti

founding team @cognition | prev: ai phd @stanford, jane street, deepmind

Stanford

Joined February 2012

Pinned
Silas Alberti
@silasalberti
Jun 4
Noticing a shift in the industry: 6 months ago everybody was tokenmaxxing. Supposedly unlimited token budgets & every CTO was afraid their team’s weren‘t using enough tokens. This is no longer the case: Our typical customer saw agent usage grow 1000% in the last few months.
Cognition
@cognition
Jun 4
AI should earn its keep. Introducing the AI Productivity Guarantee. If Devin delivers less engineering value than you’re paying for, Cognition will fund your usage until it does, up to $10 million. It’s time for the AI industry to stop maximizing tokens and start maximizing
6K
Silas Alberti
@silasalberti
Oct 7, 2025
1 trillion
1.6M
Silas Alberti
@silasalberti
Dec 28, 2022
ChatBCG: Generative AI for Slides ✨ This Christmas @josephsemrai and I finally got it working!! After DALL-E 2 for images and ChatGPT for text, the final step to make all of us redundant:  The world’s first Text-to-PowerPoint AI. 📊 ChatBCG.com 🚀
00:00
787K
Silas Alberti
@silasalberti
Apr 25, 2025
we built DeepWiki, a free encyclopedia of all GitHub repos some numbers: - 30k repos already indexed - processed 4 billion+ lines of code - the indexing alone cost $300k+ in compute spend
00:00
450K
Silas Alberti
@silasalberti
Mar 29, 2025
Wow we just ran Gemini 2.5 Pro on our evals and it got a new state of the art. Congrats to the Gemini team! Sharing preliminary results here and working on bringing it into Devin:
179K
Silas Alberti
@silasalberti
Nov 23, 2023
What could OpenAI’s breakthrough Q* be about? 1. It sounds like it’s related to Q-learning. (For example, Q* denotes the optimal solution of the Bellman equation.) 2. Alternatively, referring to a combination of the A* algorithm and Q learning. One natural guess is that it is
440K
Silas Alberti
@silasalberti
Oct 30, 2025
The era of fast coding models has begun: Cognition SWE-1.5: ~950 tok/s Cursor Composer-1: ~250 tok/s Haiku 4.5: ~140 tok/s
dax
@thdxr
Oct 29, 2025
both cursor and windsurf released models today heavily optimized for speed this is very different than the direction people have been pushing where they kick stuff off to codex for 45min but it's fast feedback loops are always what end up mattering
105K
Silas Alberti
@silasalberti
Apr 12, 2024
A new update to Devin today caused internal usage to be more than double the previous record. For the first time today Devin was the biggest contributor to the Devin repository…
185K
Silas Alberti
@silasalberti
Mar 13, 2024
Two weeks ago, I had Devin build a small SMS website summarizer and deploy it via Twilio. I was very impressed how autonomous it was. My favorite part about Devin is that it feels very collaborative. Almost like a human co-worker. My prediction is that being a strong engineer
203K
Silas Alberti
@silasalberti
Dec 1, 2022
ChatGPT is trained to not be evil. However, this can be circumvented: What if you pretend that it would actually be helpful to humanity to produce an evil response... Here, we ask ChatGPT to generate training examples of how *not* to respond to "How to bully John Doe?"
Silas Alberti
@silasalberti
Dec 20, 2024
o3 scores a 2727 ELO on Codeforces which places it 175th in the global ranking. That‘s better than ~99.9% of humans on the website (who already tend to be far above average).
50K
Silas Alberti
@silasalberti
Jun 23, 2025
I recently had a stint as a systems engineer. We built our own VM hypervisor otterlink for 10x faster startup and 200x faster snapshots (compared to EC2). One cool part was building a custom file format Blockdiff for VM disk snapshots. Sharing more in our blogpost below:
Cognition
@cognition
Jun 23, 2025
We needed instant VM snapshots for Devin but EC2 took 30+ minutes. So, @silasalberti built blockdiff—a new file format that makes snapshots 200x faster. Today, we’re open-sourcing blockdiff & sharing how it works 🔗👇
76K
Silas Alberti
@silasalberti
Oct 16, 2025
super excited to release SWE-grep and SWE-grep-mini! SWE-grep-mini achieves extreme inference speeds of >2,800 TPS: 20x faster than Haiku 4.5 while beating Sonnet 4.5, Opus 4.1 & GPT-5 on our CodeSearch eval our vision: make agentic search as fast as embedding search. the
00:28
Cognition
@cognition
Oct 16, 2025
Introducing SWE-grep and SWE-grep-mini: Cognition’s model family for fast agentic search at >2,800 TPS. Surface the right files to your coding agent 20x faster. Now rolling out gradually to Windsurf users via the Fast Context subagent – or try it in our new playground!
80K
Silas Alberti
@silasalberti
Sep 9, 2025
cognition has been ripping while maintaining the highest talent bar: since 1 year ago, revenue has grown more than 100x – but our core engineering team is still only around 30 people we're hiring exceptional people in any role but I wanted to highlight two roles in particular:
Cognition
@cognition
Sep 8, 2025
We’ve raised over $400M at a $10.2B post-money valuation to advance the frontier of AI coding agents. The round was led by Founders Fund with other existing investors including Lux, 8VC, Neo, Elad Gil, Definition Capital, and Swish VC all doubling down. We’re also joined by new
145K